AiThority Interview with Jim Kaskade, Chief Executive Officer at Conversica

Hi, Jim. Welcome to our Interview Series. Please tell us a little bit about your journey in the AI and ML technology space. How did you start at Conversica?

My personal journey began as a data engineer, working for Teradata. I later became a product manager responsible for all machine learning / deep learning and business intelligence products that realized value out of the data warehouses that Teradata powered (I worked with well-known brands like Ebay, Walmart, and Bank of America). You can say that I was born into the field of data and analytics, and have been attracted to it ever since.

My journey at Conversica began with a walk with the head of Applied AI at Google in the CTO office back in September 2019. He and I were talking about the future of AI, and specifically “applied AI”.

On that walk, I ended up looking through 3 lenses, which informed me on the future of Conversica and companies like it.

Lens #1 – ML/DL: Fun fact – I used to run a 1,000 person data science team at a company called DXC. We performed over 700 big data projects each year. During this experience, I learned that building machine learning / deep learning models was VERY hard. But what was EVEN HARDER was keeping them relevant after building them (their accuracy would decay the minute you put them into production). In today’s world of chatGPT, it’s equivalent to the information being recent or not. So the key metric for me was accuracy. And in our world today at Conversica this means accuracy in doing two things well: 1) interpreting the end-user (their intent), AND 2) taking the right next-best-action based on that interpretation (e.g. what the AI says to you in response to your question). You have to do both of these things right, together. Given our 15 years of experience doing this, we are at 98% in combined accuracy, which is unheard of in the industry. After validating the number my Google AI friend said, “you need to jump into this company tomorrow”. Which I did, obviously.

Lens #2 – Digital Transformation / Automation: Another crazy fun fact of mine. I used to run a 7,000 person team that performed 1,000 digital transformation projects each year. What I learned during this experience was that when we helped a client to become more digital, it created a BIG ISSUE – namely, when increasing digital touch between their end-customers and their business, we ended up creating an unreasonable burden on the internal human operations of the business. Think about the volume of live chat for center agents. If you were a successful ecommerce company that connected everyone on their website to a live agent, you would be investing in A LOT of bodies to answer A LOT of repetitive questions. These people-based solutions delivered poor experiences. Now try using chatbots in order to automate these responses prior to routing to a human. This digital experience was similarly frustrating to the end-customer. All we could do was recommend increasing their call-center / contact-center investments to receive more live chats, respond to emails, answer messaging application inquiries, and comment on social media channels. But that wasn’t cost effective. So you can imagine my curiosity with Conversica – powerfully human-like digital agents that provide the necessary “buffer” between the massive amount of digital touch and human workforces. The focus on AI automation for marketing, sales, and customer success makes a ton of sense, especially IF you could make your AI powerfully human-like, which Conversica can.

Lens #3 – Scaling for Enterprise Businesses:  An AI startup can’t be trusted with mid-market and enterprise customer candidates, right? If you are Oracle, AT&T, or similar Fortune 1000, your customers are too important to hand to an AI – especially one that generates inaccurate responses, goes rouge, or worse can’t respond because it wasn’t programmed for the question being presented. We now know that 4 out of 5 end-users simply give up on the conversation because of the lack of immediacy, relevancy, and overall accuracy.  Well, my hike with my Google AI friend led to my understanding of Google PaLM, Meta’s LLaMA and, of course Microsoft and OpenAI’s GPT. But the history of Google’s work with technologies like GPT, seemed well ahead of Microsoft and OpenAI, at the time. I was very intrigued, and this walk ultimately led to my early direction for Conversica’s use of GPT in 2020. We knew that we needed to provide a “mature” version of large language model technology – something “Enterprise ready”.

Just recently, we disclosed our enterprise features of our GPT-powered chat solution, called Conversica Chat. It leverages the innovation from AI powerhouses Microsoft, Google, and Meta.

Conversica Chat uniquely applies the best of generative AI technology and then ads enterprise features to engage website visitors in humanlike, two-way conversations that drive revenue growth outcomes. But webchat is just one piece of our full-lifecycle solution; our technology can carry the conversation across many communication channels (chat, email, sms, messaging, social, and even voice) and throughout the customer journey (from marketing to sales and customer success.)

Recommended: AiThority Interview with Andy Pitre, Executive Vice President of Product at HubSpot

Please tell us more about your role in the conversation automation industry. How do you define new benchmarks for your product and marketing team?

We are the largest commercial provider of Conversation Automation Applications for the Enterprise in the industry focusing on revenue acceleration. We’re constantly pushing the boundaries of what’s possible with AI-powered conversational technology, and we’re always looking for new ways to help businesses automate their marketing, sales, and customer success processes.

There is a specific benchmark, unique to us at this point. We call it “Conversation Qualified”. It’s a measure of the number of end-users our AI reaches out to and has a two-way conversation to the number of sales-ready opportunities handed over to your revenue teams.

As a marketeer, you care about MQLs – how many inbound marketing-qualified leads can you pass to sales development teams? Well, we take the MQL and turn that into a CQL (Conversation  Qualified Lead) from Marketing. Instead of an open rate, click rate, or lead scored lead, we simply ask, “Did the lead have a conversation with the AI?” – meaning an actual two-way exchange? If so, then we believe that it’s an MQL worth considering. But wait, even better, the AI will NOT pass that lead to a sales development or business development person until it has qualified it as “sales ready”. Oh My God!!! This is a sales person’s dream!

As a sales person, our AI further qualifies the lead through two-way communications. So you might be looking at SQLs as your main metric. Well, again, we take the SQL and turn that into a CQL from Sales. Instead of a Sales Qualified Lead where the prospect is willing to take a meeting with a sales person, we further qualify that lead to make sure it’s worthy of the sales representative. Further two-way conversations occur asking qualifying questions. The AI will NOT pass that lead to a sales rep until it has qualified it as “sales ready”, but this time at an even more qualified level.

As a customer success person, our AI further qualifies upsell and cross sell opportunities. We hand opportunities over to the Customer Success Manager or sales rep on installed base growth opportunities, after they have been conversation qualified.

To define new benchmarks for our product teams, it really comes down to how well our messaged to conversation qualified rates perform at various stages of lead maturity. Our customer success AI, for example, is trying to cross-sell and upsell to existing customers.

Ultimately, our goal is to provide the best possible experience for the end-user, when engaging the digital brand. For our client brands, it’s all about revenue. How much MORE revenue can we help generate by using AI automation.

In the era of ChatGPT and generative AI, how does Conversica continue to lead the market with its breakthrough AI? Tell us more about the AI engine at the backdrop of your recent AI developments?

Conversica has been pioneering the field of conversational AI since 2007.

One of the key ways that we differentiate ourselves from other AI solutions is our focus on driving outcomes for specific revenue-centric business use cases. We believe that conversational AI can provide tremendous value to businesses, but it requires a specialized approach that takes into account the unique needs of revenue teams including marketing, sales, and customer success — which are the target for the solutions we create.

To that end, we’ve developed a proprietary AI engine that is specifically designed for business outcomes. Our AI engine is built on a foundation of natural language processing, machine learning, and advanced technologies like GPT, but what enables the most value is the fact that  it has been trained on over a billion of real-world conversations to ensure that it’s highly accurate and effective in getting end-customers to buy products and services.

How do we stay ahead of the “state of the art”? We adopt improvements quickly, moving from GPT 3.5 to 4, for example. But we also add value deploying Enterprise-class features that ensure our customers’ brands are protected.

Recommended: AiThority Interview with Malcolm Koh, Director, CX Practice at Zendesk

What does your ideal customer profile look like? What kind of customers are you planning to reach with your new AI products?

Our ideal customer profile is any business that is looking to automate its sales and marketing processes and engage with its customers in a more personalized and effective way in a scalable manner. We work with a wide range of businesses, from medium to large enterprises, in a variety of industries including technology, financial services, sports and entertainment, automotive, and more.

One of the key benefits of our AI-powered conversation automation solutions is that they can be tailored to meet the specific needs of each individual customer. We have hundreds of conversations optimized over more than a billion interactions, all of which are easily customizable to fit a wide variety of use cases and industries.

With our new chat feature, we’re focused on reaching even more businesses that are looking to provide a better webchat experience and capture more revenue opportunities without adding work for their employees. Conversica Chat allows businesses to thread that needle. It’s a personalized, tailored, humanlike experience for the end user that autonomously drives conversations toward business goals.

The AI handles the whole interaction to gather information, answer questions, provide recommendations, schedule meetings—and it can carry all that over to a different channel if the contact wants to talk by email or SMS, for example. Human employees only need to get involved when the prospect or customer is ready for the next step in their journey.

ChatGPT conversations are everywhere. How do you see ChatGPT and other generative AI apps playing a larger role in your industry?

At Conversica, we believe that ChatGPT and other generative AI applications will have a significant impact across industries. Technologies like Conversica Chat, which we just relaunched, can revolutionize the way businesses interact with customers by providing a more human-like and personalized experience.

Access to large language models will support the development of many more innovative applications beyond chatGPT. Look at us, we’ve created an Enterprise-version of chatGPT that also drives users to purchase more products and services. Our application of generative AI empowers revenue teams to have more meaningful conversations with customers at scale, across all communication channels, driving revenue growth.

How to make AI innovations more inclusive and equitable?

Making AI innovations more inclusive and equitable is an important issue for the entire industry. Our AI takes on personas with both gender and ethnic diversity. Also, with generative AI, we have to remove biases that naturally impact equity and inclusion. We recognize that the data we use to train our AI models can introduce biases, fake news and even hate speech, which can perpetuate inequities.

To address this, we use a variety of techniques, including diversifying our training data and implementing algorithms that detect and correct for bias. We have filters on the input and output of our AI to remove bias. We build “brand-safe” models specific for each of our clients that guarantee that your brand is represented better than your best employee. Additionally, we support initiatives that promote the responsible use of AI and advocate for policies that ensure equitable applicatoin of AI technologies. By working together as an industry, we can develop AI innovations that benefit everyone, safely, creating a more inclusive and equitable world.

Recommended: AiThority Interview with Ben Plomion, Chief Marketing Officer at Dibbs

Your take on AI ethics and democratization of data science ecosystems?

AI ethics and the democratization of data science ecosystems are pivotal issues that need to be addressed as AI becomes more widely used. We have created a “Responsible AI Framework” which guides us, our partners, and our customers. We believe that AI must be developed responsibly with consideration given to ethical implications. We provide guidance on topics like disclosing your AI to end-customers.

We also collaborate and share knowledge to the industry to advance the responsible use of AI. By promoting ethics and democratization, we can create technology that benefits everyone and promotes a more just and equitable society.

Thank you, Jim! That was fun and we hope to see you back on AiThority.com soon.